Towards the Tactile Internet: Feature Similarity-Driven Virtualized MEC Slicing

Abstract

Network slicing and mobile edge computing (MEC) have long been discussed by both academia and industry, owing to their ability to enable diverse, cost-effective scenarios in 5th generation mobile communication technology wireless networks. The collaborative term of MEC as multi-access edge computing supports the intelligent utilization of network, computing, and storage resources for efficient resource management and service availability. However, ensuring service availability in a tactile internet environment is still challenging, while parallel and dynamic resource utilization becomes the key issue. Virtual resource management and the slicing approach can provide dynamically classified scenarios by arranging network entities and resources into multiple logically independent networks. Hence, the independent virtual functions through MEC-based access network slicing can improve existing MEC approaches for tactile internet services. This manuscript proposes research on centralized virtual slicing to efficiently handle distributed MEC resources, including computing, networking, and storage. Besides, these resources can be efficiently utilized in the proposed MEC slicing approach, avoiding over-provisioning of available MEC resources. Whereas over-provision of resources greatly impacts the high bandwidth and ultra-low latency requirements of tactile services. A sliced MEC (SL-MEC) framework is proposed, considering the resource coupling throughout the similarity matrix. With respect to tactile internet, the SL-MEC generates the aggregated slices 6× faster than the benchmark algorithms, with guaranteed resource availability.